Sam Tobin-Hochstadt

It needs three things. (1) To fix whatever is causing it to segfault currently. (2) To enable it with a build-time or run-time option instead of `if (1)`. (3) To...

See also https://chromium.googlesource.com/v8/v8/+/refs/heads/master/src/diagnostics/perf-jit.cc

Implementing this for Racket CS shouldn't be that hard -- the key thing is to figure out where the machine code gets written, and add something to write to the...

More information about formats and integration with perf: https://bugzilla.mozilla.org/show_bug.cgi?id=1759173 https://github.com/torvalds/linux/blob/816e51dfb5ba47e4f989af656c956a8c0cc686c0/tools/perf/Documentation/jitdump-specification.txt

@Aeva it would be helpful if you described the packaging scenarios you're thinking about so as to think precisely about what files get included. Currently, the "math-win32-x86_64" package contains only...

For a simple example package, what sort of metadata are you thinking of? I think the output of `raco pkg new` contains all the metadata I think of as usual.

> Sorry, which part did you want a clarification on? What things you're doing with racket to generate what kinds of artifacts, and what you want to ensure about them....

As for Racket providing a tool as part of the standard distribution, I'm not personally in favor for a couple reasons. One is just that it seems easy and effective...

@Aeva For your first kind of sample repository, `raco pkg new` produces exactly that. I think that https://github.com/racket-templates/package is similar. For the second kind, that seems quite specific to have...

Thanks for the list, @Aeva. I will leave aside the future static-linking possibilities, since those seem far in the future and it remains to be seen what that would mean...